Linguistic Areas Lyle Campbell University of Hawai’i Manoa Areal linguistics is about the diffusion of structural features across the languages of a geographical area. The term linguistic area refers to a geographical area in which, due to borrowing and language contact, languages of a region have come to share structural features – Terms: Sprachbund, diffusion area, convergence area, adstratum. The Balkans -- the best known linguistic area. Languages: Greek, Albanian, Serbo-Croatian, Bulgarian, Macedonian, and Romanian (some add Romani and Turkish). x x Balkans areal traits: 1. Central vowel /ɨ/ (or /ə/) (not in Greek or Macedonian); 2. Syncretism of dative and genitive cases (merged in form and function). Romanian fetei ‘to the girl’ or ‘girl’s’: am data o carte fetei ‘I gave the letter to the girl’ and frate fetei ‘the girl’s brother’; 3. Postposed articles (not in Greek). Bulgarian məʒət ‘the man’, məʒ ‘man’. 4. Periphrastic future (with auxiliary corresponding to ‘want’ or ‘have’, not in Bulgarian or Macedonian). Romanian voi fuma ‘I will smoke’ (literally ‘I want (to) smoke’) and am a cínta ‘I will sing’ (literally ‘I have sing’); 5. Periphrastic perfect (with auxiliary corresponding to ‘have’); 6. Absence of infinitives (constructions equivalent to ‘I want that I go’ for ‘I want to go’); 7. Double marking of animate objects by pronoun copy. Romanian i-am scris lui Ion ‘I wrote to John’, lit. ‘to.him-I wrote him John’, and Greek ton vlépo ton jáni ‘I see John’, literally ‘him.Acc I see him.Acc John’. South Asia (Indian sub-continent) Linguistic Area Languages: Indo-Aryan (branch of Indo-European), Dravidian, Munda, Tibeto-Burman (branch of Sino- Tibetan) families. South Asian Areal traits: 1. Retroflex C’s, particularly retroflex stops; 2. Absence of prefixes (accept in Munda); 3. Dative-subject construction (dative-experiencer. Hindi mujhe maaluum thaa ‘I knew it’ [mujhe ‘to me’ + know + Past]); 4. Subject-Object-Verb (SOV) word order, postpositions; 5. Absence of a verb ‘to have’; 6. ‘Conjunctive or absolutive participles’ (subordinate with non- finite verbs (participles), preposed; e.g. relative clauses precede the nouns the modify; 7. Morphological causatives (verbal suffix); South Asian Areal traits: 1. Retroflex C’s, particularly retroflex stops; 2. Absence of prefixes (accept in Munda); 3. Dative-subject construction (dative-experiencer. Hindi mujhe maaluum thaa ‘I knew it’ [mujhe ‘to me’ + know + Past]); 4. Subject-Object-Verb (SOV) word order, postpositions; 5. Absence of a verb ‘to have’; 6. ‘Conjunctive or absolutive participles’ (subordinate with non- finite verbs (participles), preposed; e.g. relative clauses precede the nouns they modify; 7. Morphological causatives (verbal suffix); South Asia (the Indian sub-continent) linguistic area (cont.) 8. ‘Explicator compound verbs’ (auxiliary from a limited set completes the sense of the preceding main verb, the two verbs together refer to a single event. Hindi le jaanaa ‘to take (away)’ [‘take’ + ‘go’]); 9. Sound symbolic forms based on reduplication, often with k suffixed (Kota [Dravidian]: kad-kadk ‘[heart ] beats fast with guilt or worry’; a:nk-a:nk ‘to be very strong [of man, bullock], very beautiful [of woman]’). Some of these features are not limited to the Indian sub- continent, are also in neighboring languages (e.g. SOV word order is in much of Eurasia and N Africa). Some traits are not independent of one another, e.g. languages with SOV word order tend to have non-finite (participial) subordinate clauses, and not to have prefixes. Mesoamerica Languages: Nahua (branch of Uto-Aztecan), Totonacan, Otomanguean, Mixe-Zoquean, Mayan, Xinkan, Tarascan, Cuitlatec, Tequistlatecan, Huave. Mesoamerican areal traits: Five areal traits are shared by nearly all Mesoamerican languages, but not by neighboring languages beyond this area, and these are considered particularly diagnostic of the linguistic area: 1. Nominal possession: his-dog the man ‘the man’s dog’. Pipil (Uto-Aztecan): i-pe:lu ne ta:kat, lit. ‘his-dog the man’; 2. Relational nouns (locative expressions composed of noun roots +possessive pronominal affixes), my-head for ‘on me’. Tz’utujil (Mayan): č-r-i:x ‘behind it, in back of it’: č- ‘at, in’, r- ‘his/her/its’, and i:x ‘back’; č-w-i:x ‘behind me’, lit. ‘at-my-back’; 3. Vigesimal numeral systems . Chol (Mayan): hun-k’al ‘20’ (1x20), čaʔ-k’al ‘40’ (2x20), uš-k’al ‘60’ (3x20), hoʔ-k’al ‘100’ (5x20), hun-bahk’ ‘400’ (1-bahk’ ), čaʔ-bahk’ ‘800’ (2x400); 4. Non-verb-final word order (no SOV) – Mesoamerica is surrounded SOV languages; Mesoamerica (cont.) 5. a number of calques (loan translation compounds) shared by Mesoamerican languages:‘boa’ = ‘deer-snake’, ‘egg’ = ‘bird-stone/bone’, ‘lime’ = ‘stone(-ash)’, ‘knee’ = ‘leg-head’, and ‘wrist’ = ‘hand-neck’. Since these five traits are shared almost unanimously throughout Mesoamerican languages but are generally not found in the languages just beyond the borders of Mesoamerica, these traits are strong evidence supporting the Mesoamerica as a linguistic area. Four of these five traits have essentially the same distribution, clustering at the borders of Mesoamerica. Such bundling is uncommon in linguistic areas. A large number of other features are shared among several Mesoamerican languages, but some are not found in all the languages of the area, while others are found also in languages beyond the borders of the area (see Campbell et al. 1986.) The Northwest Coast Linguistic Area Languages: Tlingit, Eyak, the Athapaskan languages of the region, Haida, Tsimshian, Wakashan, Chimakuan, Salishan, Alsea, Coosan, Kalapuyan, Takelma, Lower Chinook. NW Coast areal traits: Elaborate systems of C’s: series of glottalized stops and affricates, labiovelars, multiple laterals, uvular stops; typically few labials (completely lacking in Tlingit and Tillamook, very limited in Eyak and most Athabaskan languages); several uvular consonants. Few V’s, only 3 (i, a, o, or i, a, u ) in several languages, 4in others. Several have pharyngeals (ħ, ʕ), most have glottalized resonants and continuants. Shared morphological traits: suffixes, near absence of prefixes; reduplication (of several sorts, signaling iteration, continuative, progressive, plural, collective, distribution, repetition, diminutive, etc.); numeral classifiers; alienable/inalienable oppositions in nouns; evidential markers in verbs; verbal locative- directional markers. NW linguistic area (cont.) Masculine/feminine gender (shown in demonstratives and articles); visibility/invisibility opposition in demonstratives. Aspect is more important than tense. All except Tlingit have passive-like constructions. The negative appears as the first element in a clause regardless of the usual word order. Overt marking of nominal plurals is absent or limited, as in many American Indian languages. Northwest Coast languages also have lexically paired suppletive singular and plural verb stems (that is, a lexical root may be required with a plural subject which is entirely different from the root used with a singular subject). Some other traits are shared by a smaller number of Northwest Coast languages. HOW LINGUISTIC AREAS ARE DEFINED Criteria that have been used: (1) The number of traits shared by languages in an area, (2) Bundling of the traits in some significant way (e.g. at the same geographical boundaries), (3) The weight or complexity of different areal traits (some are accorded more significance for determining areal affiliation on the assumption that they are more difficult to acquire by diffusion than others). To establish a LA, the more shared features, the better. LAs in which many diffused traits are shared among the languages are considered better established. Difference of opinion about how many traits it takes. Some believe even one shared trait is enough to define a weak linguistic area. In the best known linguistic areas, the number and kind of shared traits vary a lot. The idea that greater weight should be attributed to some traits for defining linguistic areas – example: borrowed word order in Ethiopian linguistic area. Ethiopian Semitic languages interconnected word-order patterns, borrowed from neighboring Cushitic languages. Several of these traits reflect the diffusion of the SOV (Subject-Object- Verb) basic word order typology of Cushitic languages into the formerly VSO Ethiopian Semitic languages. Typologically the orders Noun-Postposition, Verb-Auxiliary, Relative Clause-Head Noun, and Adjective-Noun are all correlated and tend to co-occur with SOV order cross-linguistically. How to count: several shared traits?, or one general typological trait (SOV typology)? If one, it ranks high for defining a LA, since it is much more difficult for a language to change so much of its basic structure under areal influence than it is to acquire less complex traits. Bundling of areal traits, clustering at the boundaries of a linguistic area, might be required for defining linguistic areas, though this has proven a poor criterion. Linguistic areas are similar to traditional dialects, where often one trait spreads across more territory than another trait, so that their boundaries (or territories) do not coincide (do not ‘bundle’). Typically the geographical extent of individual traits may vary considerably. However, in the rare situation where the traits do coincide at a clear boundary, the definition of a linguistic area matching their boundary is relatively secure (e.g. Mesoamerica). Problem: Lack of bundling raises serious questions for the definition of linguuistic